FFmpeg/libavcodec/riscv/bswapdsp_rvb.S

/*
 * Copyright © 2022 Rémi Denis-Courmont.
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#include "libavutil/riscv/asm.S"
#include "libavutil/riscv/bswap_rvb.S"

#if (__riscv_xlen >= 64)
func ff_bswap32_buf_rvb, zbb
        lpad    0
        slli    a2, a2, 2
        bswap32_rvb a0, a1, a2
endfunc
#endif
lavc/bswapdsp: RISC-V B bswap_buf Simply taking the Zbb REV8 instruction into use in a simple loop gives some significant savings: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 771.0 But we can also use the 64-bit REV8 as a pseudo-SIMD instruction with just one additional shift, and one fewer load, effectively doubling the bandwidth. Consequently, this patch is useful even if the compile-time target has Zbb enabled for C code: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 341.0 (this patch) On the other hand, this approach fails miserably for bswap16_buf as the ratio of shifts and stores becomes unfavorable compared to naïve C: bswap16_buf_c: 1542.0 bswap16_buf_rvb_b: 1803.7 Unrolling to process 128 bits (4 samples) at a time actually worsens performance ever so slightly: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 408.5 2 years ago			`/*`
			`* Copyright © 2022 Rémi Denis-Courmont.`
			`*`
			`* This file is part of FFmpeg.`
			`*`
			`* FFmpeg is free software; you can redistribute it and/or`
			`* modify it under the terms of the GNU Lesser General Public`
			`* License as published by the Free Software Foundation; either`
			`* version 2.1 of the License, or (at your option) any later version.`
			`*`
			`* FFmpeg is distributed in the hope that it will be useful,`
			`* but WITHOUT ANY WARRANTY; without even the implied warranty of`
			`* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU`
			`* Lesser General Public License for more details.`
			`*`
			`* You should have received a copy of the GNU Lesser General Public`
			`* License along with FFmpeg; if not, write to the Free Software`
			`* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA`
			`*/`

			`#include "config.h"`
			`#include "libavutil/riscv/asm.S"`
riscv: factor out the bswap32 assembler 1 year ago			`#include "libavutil/riscv/bswap_rvb.S"`
lavc/bswapdsp: RISC-V B bswap_buf Simply taking the Zbb REV8 instruction into use in a simple loop gives some significant savings: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 771.0 But we can also use the 64-bit REV8 as a pseudo-SIMD instruction with just one additional shift, and one fewer load, effectively doubling the bandwidth. Consequently, this patch is useful even if the compile-time target has Zbb enabled for C code: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 341.0 (this patch) On the other hand, this approach fails miserably for bswap16_buf as the ratio of shifts and stores becomes unfavorable compared to naïve C: bswap16_buf_c: 1542.0 bswap16_buf_rvb_b: 1803.7 Unrolling to process 128 bits (4 samples) at a time actually worsens performance ever so slightly: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 408.5 2 years ago
			`#if (__riscv_xlen >= 64)`
lavu/riscv: count bytes rather than words for bswap32 This removes the dependency on Zba at essentially zero cost. 5 months ago			`func ff_bswap32_buf_rvb, zbb`
lavc/riscv: add forward-edge CFI landing pads 5 months ago			`lpad 0`
lavu/riscv: count bytes rather than words for bswap32 This removes the dependency on Zba at essentially zero cost. 5 months ago			`slli a2, a2, 2`
riscv: factor out the bswap32 assembler 1 year ago			`bswap32_rvb a0, a1, a2`
lavc/bswapdsp: RISC-V B bswap_buf Simply taking the Zbb REV8 instruction into use in a simple loop gives some significant savings: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 771.0 But we can also use the 64-bit REV8 as a pseudo-SIMD instruction with just one additional shift, and one fewer load, effectively doubling the bandwidth. Consequently, this patch is useful even if the compile-time target has Zbb enabled for C code: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 341.0 (this patch) On the other hand, this approach fails miserably for bswap16_buf as the ratio of shifts and stores becomes unfavorable compared to naïve C: bswap16_buf_c: 1542.0 bswap16_buf_rvb_b: 1803.7 Unrolling to process 128 bits (4 samples) at a time actually worsens performance ever so slightly: bswap_buf_c: 1081.0 bswap_buf_rvb_b: 408.5 2 years ago			`endfunc`
			`#endif`